An Investigation of University Selection Procedures

This article re-examines the selection process in the context of a sample of applicants to read Philosophy, Politics and Economics at Oxford University in 1978. We have observed that a stable equation can be estimated that correctly classifies a high proportion of candidates between accepted and rejected categories and forecasts well out of sample. Performance in Oxford's own entrance exam was found to be the most important public examination influence on the probability of being accepted. A- and O-levels, while less influential, were still highly significant. Inclusion of individual entrance and A-level marks substantially improved the performance of the function and some exam results were observed to be more significant than others. Socioeconomic variables as a group were highly significant. While women had a slightly higher probability of being admitted ceteris paribus than men there was not much evidence of sex discrimination. Far more important was the type of school from which the candidate came. Finally, in contrast to selection, the decision to award a scholarship or exhibition appears to be based virtually entirely on entrance exam performance and, to a lesser extent, A-level grades.